First report on physician assessment and clinical acceptability of custom-retrained AI models for clinical target volume and organs-at-risk auto-delineation for post-prostatectomy patients.

Researchers

Journal

Modalities

Models

Abstract

To assess the clinical acceptability of a commercial deep-learning-based auto-segmentation (DLAS) prostate model that was re-trained using institutional data foe delineation of the clinical target volume (CTV) and organs-at-risk (OARs) for post-prostatectomy patients, accounting for clinical and imaging protocol variations.CTV and OARs of 109 prostate-bed patients were utilized to evaluate the performance of the vendor-trained (VT) model and custom re-trained DLAS models using different training quantities. Two new models for OAR structures were re-trained (n = 30, 60 datasets), while separate models were trained for a new CTV structure (n = 30, 60, 90 datasets), with the remaining datasets used for testing (n = 49, 19). The dice similarity coefficient (DSC), Hausdorff distance (HD95), and mean surface distance (MSD) were evaluated. Six radiation oncologists performed a qualitative evaluation scoring both preference and clinical utility for blinded structure sets. Physician consensus (PC) datasets identified from the qualitative evaluation were utilized towards a separate CTV model.Both the 30- and 60-case re-trained OAR models had median DSC values between 0.91-0.97, improving significantly over the VT-model for all OARs except the penile bulb. The brand new 60-case CTV model had a median DSC of 0.70 improving significantly over the 30-case model. DLAS (60-case model) and manual contours were blinded and evaluated by physicians with contours deemed acceptable or precise for 87% and 94% of cases for DLAS and manual delineations, respectively. DLAS-generated CTVs were scored precise or acceptable in 54% of cases, as compared to the manual delineation value of 73%. The 30-case PC CTV model did not show a significant difference compared to the randomly selected models.Custom re-training using institutional data leads to performance improvement in the clinical utility and accuracy of DLAS for post-prostatotomy patients. A small number of datasets are sufficient for building an institutional site-specific DLAS OAR model, as well as for training new structures. Data indicates the workload for identifying training datasets could be shared among groups for the male pelvic region, making it accessible to clinics of all sizes.Copyright © 2023. Published by Elsevier Inc.